from sklearn.impute import SimpleImputer
import ipywidgets as widgets
import matplotlib.pyplot as plt
upload = widgets.FileUpload(accept='.csv,.xlsx,.json', multiple=False)
def process_file(upload):
uploaded_file = next(iter(upload.value))
file_name = uploaded_file['name']
content=uploaded_file['content']
file_data = io.BytesIO(content)
if file_name.endswith('.csv'): df = pd.read_csv(file_data)
elif file_name.endswith('.json'): df = pd.read_json(file_data)
elif file_name.endswith('.xlsx'): df = pd.read_excel(file_data, engine='openpyxl')
print('unsupported file format') print('error loading file')analyze_button = widgets.Button(description="Process File")
output = widgets.Output()
data=process_file(upload)
numeric= data.select_dtypes(include=[np.number]).columns
categoric=data.select_dtypes(exclude=[np.number]).columns
print('number of numeric columns:',len(numeric)) print(' number of categorical columns:',len(categoric)) print(' '.join(['_' for i in range(100)])) print('number of missing values in each column:') print(data.isnull().sum())
print(' '.join(['_' for i in range(100)])) print('treating the missing values:') if data[i].isnull().sum() >= (1/4) * len(data):
columns_to_drop.append(i)
if pd.api.types.is_numeric_dtype(data[i]):
imputer = SimpleImputer(missing_values=np.nan, strategy='mean')
imputer = SimpleImputer(missing_values=np.nan, strategy='most_frequent')
data[i] = imputer.fit_transform(data[[i]]).ravel()
data=data.drop(columns=columns_to_drop)
print(data.isnull().sum())
print(' '.join(['_' for i in range(100)])) print('statestical information:') print(' '.join(['_' for i in range(100)])) print('data visualization:') numeric_cols=data.select_dtypes(include=[np.number]).columns
num_cols=len(numeric_cols)
print("Not enough numeric columns for correlation subplots.") plt.figure(figsize=(14,10))
sns.heatmap(data[numeric_cols].corr(), annot=True, cmap='coolwarm', linewidths=0.5)
plt.title('Correlation Heatmap') print(' '.join(['_' for i in range(100)])) print('pairwise correlation:') print("Not enough numeric columns for pairwise plots.") fig, axes = plt.subplots(nrows=num_cols, ncols=num_cols, figsize=(6*num_cols, 6*num_cols))
fig.suptitle('pairwise correlation scatter plots', fontsize=16) for (i, j) in itertools.product(range(num_cols), range(num_cols)):
axes[i, j].hist(data[numeric_cols[i]], bins=20, color='skyblue', edgecolor='black')
axes[i, j].set_title(f'{numeric_cols[i]} (Histogram)') axes[i, j].scatter(data[numeric_cols[j]], data[numeric_cols[i]], alpha=0.5, c='green')
axes[i, j].set_xlabel(numeric_cols[j])
axes[i, j].set_ylabel(numeric_cols[i])
plt.tight_layout(rect=[0, 0, 1, 0.96])
p=analyze_button.on_click(on_button_click)
display(analyze_button, output)
